8P22 MTUS1 Gene Product ATIP3 Is a Novel Anti-Mitotic Protein Underexpressed in Invasive Breast Carcinoma of Poor Prognosis

8p22 MTUS1 Gene Product ATIP3 Is a Novel Anti-Mitotic Protein Underexpressed in Invasive Breast Carcinoma of Poor Prognosis

Sylvie Rodrigues-Ferreira1, Anne Di Tommaso1, Ariane Dimitrov2, Sylvie Cazaubon1, Nade`ge Gruel3,4, He´le`ne Colasson1, Andre´ Nicolas5, Nathalie Chaverot1, Vincent Molinie´ 6, Fabien Reyal2, Brigitte Sigal-Zafrani4,5, Benoit Terris7, Olivier Delattre3,4, Franc¸ois Radvanyi2, Franck Perez2, Anne Vincent-Salomon4,5, Clara Nahmias1* 1 Institut Cochin, Universite´ Paris Descartes, Inserm U567, CNRS UMR8104, Paris, France, 2 CNRS UMR144, Institut Curie, Paris, France, 3 Inserm U830, Institut Curie, Paris, France, 4 Translational Research Department, Institut Curie, Paris, France, 5 Pathology Department, Hopital Curie, Paris, France, 6 Pathology Department, Hopital St. Joseph, Paris, France, 7 Pathology Department, Hopital Cochin, Paris, France

Abstract

Background: Breast cancer is a heterogeneous disease that is not totally eradicated by current therapies. The classification of breast tumors into distinct molecular subtypes by gene profiling and immunodetection of surrogate markers has proven useful for tumor prognosis and prediction of effective targeted treatments. The challenge now is to identify molecular biomarkers that may be of functional relevance for personalized therapy of breast tumors with poor outcome that do not respond to available treatments. The Mitochondrial Tumor Suppressor (MTUS1) gene is an interesting candidate whose expression is reduced in colon, pancreas, ovary and oral cancers. The present study investigates the expression and functional effects of MTUS1 gene products in breast cancer.

Methods and Findings: By means of gene array analysis, real-time RT-PCR and immunohistochemistry, we show here that MTUS1/ATIP3 is significantly down-regulated in a series of 151 infiltrating breast cancer carcinomas as compared to normal breast tissue. Low levels of ATIP3 correlate with high grade of the tumor and the occurrence of distant metastasis. ATIP3 levels are also significantly reduced in triple negative (ER- PR- HER2-) breast carcinomas, a subgroup of highly proliferative tumors with poor outcome and no available targeted therapy. Functional studies indicate that silencing ATIP3 expression by siRNA increases breast cancer cell proliferation. Conversely, restoring endogenous levels of ATIP3 expression leads to reduced cancer cell proliferation, clonogenicity, anchorage-independent growth, and reduces the incidence and size of xenografts grown in vivo. We provide evidence that ATIP3 associates with the microtubule cytoskeleton and localizes at the centrosomes, mitotic spindle and intercellular bridge during cell division. Accordingly, live cell imaging indicates that ATIP3 expression alters the progression of cell division by promoting prolonged metaphase, thereby leading to a reduced number of cells ungergoing active mitosis.

Conclusions: Our results identify for the first time ATIP3 as a novel microtubule-associated protein whose expression is significantly reduced in highly proliferative breast carcinomas of poor clinical outcome. ATIP3 re-expression limits tumor cell proliferation in vitro and in vivo, suggesting that this protein may represent a novel useful biomarker and an interesting candidate for future targeted therapies of aggressive breast cancer.

Citation: Rodrigues-Ferreira S, Di Tommaso A, Dimitrov A, Cazaubon S, Gruel N, et al. (2009) 8p22 MTUS1 Gene Product ATIP3 Is a Novel Anti-Mitotic Protein Underexpressed in Invasive Breast Carcinoma of Poor Prognosis. PLoS ONE 4(10): e7239. doi:10.1371/journal.pone.0007239 Editor: Mikhail V. Blagosklonny, Roswell Park Cancer Institute, United States of America Received June 5, 2009; Accepted September 5, 2009; Published October 1, 2009 Copyright: ß 2009 Rodrigues-Ferreira et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by the Centre National de la Recherche Scientifique (CNRS), the Institut national de la sante et de la recherche medicale (Inserm), the Association pour la Recherche contre le Cancer (ARC), the Ligue Nationale Contre le Cancer (LNCC) comite Ile-de-France and the Institut National du cancer (INCa). S.R-F was supported by an ARC fellowship and A.DiT by an INCa/Inserm fellowship. The funders had no role in study design, data collectionand analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected]

Introduction Over the past decade, traditional prognostic and predictive biomarkers including estrogen receptor (ER), progesterone Despite extensive progress in the field of breast cancer, this receptor (PR) and human epidermal growth factor receptor 2 disease remains the leading cause of death by malignancy in (HER2) have been routinely assessed by immunohistochemistry in women worldwide. The identification of novel molecular markers newly diagnosed breast cancer [1]. Tumors that are positive for associated with poor clinical outcome of the disease is of ER or PR are generally treated by hormonal therapy whereas considerable interest for tumor classification and treatment, and patients with HER2 positive tumors are candidates for targeted represents an issue of major public health importance. anti-HER2 therapies. Tumors exhibiting a triple negative

PLoS ONE | www.plosone.org 1 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer phenotype (negative for ER, PR and HER2) currently have no were obtained from breast mammoplasty surgery. Total RNA was targeted treatment available and remain of poor prognosis [2]. extracted by the cesium chloride method and hybridized to DNA The development of gene profiling studies has proven microarrays (Human Genome, Affymetrix HG-U133 set) as particularly powerful in classifying breast tumors into distinct previously described [22]. Microarray data reported here are in molecular subgroups with different biological characteristics and accordance with MIAME guidelines and have been partially clinical outcome [3]. Three main ‘‘intrinsic’’ sub-types of tumors published elsewhere [22]. All samples were reviewed and classified (luminal, HER2-like and basal-like) have been distinguished that by the same breast pathologist (AVS). Histologic nuclear grades turned out to be similar to, although not just recapitulating, those were scored (I to III) according to Elston and Ellis [23]. Estrogen defined by clinical practice [4–6]. Thus, the luminal group of receptor (ER) and progesterone receptor (PR) expression were tumors is characterized by ER/PR expression whereas HER2-like determined by immunohistochemistry [24] and were considered tumors are mainly ER negative and strongly HER2 positive. negative when less than 10% of the tumor cell nuclei showed Basal-like carcinomas constitute a heterogeneous subtype of staining. HER2 expression was evaluated by immunostaining as tumors with agressive clinical behavior, that predominantly lack described previously [24] and was considered positive when IHC the expression of surrogate markers ER, PR and HER2 [2]. staining was 3+ (uniform, intense membrane staining of more than Recent high-throughput, genome-wide studies have provided an 30% of invasive tumor cells), as recommended by Wolff et al. [25]. integrative analysis of genomic copy number alterations and This study was approved by the institutional review boards and transcriptomic profiles of breast cancer [7,8], that may facilitate ethics committee of the Institut Curie. the identification of oncogenic and tumor suppressor pathways specifically associated with each breast cancer molecular subtype Cell culture, plasmid constructs and transfections [9,10]. A considerable number of genetic elements of interest Human cancer cell lines MDA-MB-231, MDA-MB-468, SK- altered in breast cancer have thus been identified [11]. The MES, HeLa and HeLa-H2B were maintained at 37uC with 5% challenge now is to functionally validate the products of these CO2 in DMEM 4.5 g/l glucose supplemented with 10% fetal calf genes to determine whether they may be used clinically as novel serum (FCS). MCF7 breast cancer cells were grown in DMEM/ prognostic or predictive biomarkers of breast tumors with poor F12 medium with 5% FCS. All cancer cell lines are available from outcome. ATCC. HeLa-H2B cells stably expressing the mCherry-histone The mitochondrial tumor suppressor (MTUS1) gene localizes at were kindly provided by Dr. Vale´rie Doye (Institut Jacques 8p22, a chromosomal region frequently deleted in epithelial Monod, Paris). The source, tumor type and histologic character- cancers including in poor prognosis breast cancer [12]. Previous istics of breast cancer cell lines used in this study have been studies have shown that MTUS1 expression levels are reduced in described in details elsewhere [26,27]. cancers from pancreas [13], ovary [14], colon [15] and head-and- The coding sequence of human ATIP3a (3812 bp) was obtained MTUS1 neck [16]. A deletion polymorphism spanning a coding by ligating partial 39 ATIP sequence isolated from human lung exon [17] has been shown to significantly associate with decreased cDNA library [19] with the 59 region of IMAGE clone ID risk of familial breast cancer [18]. Our group has previously 3835953 (ATCC). Sequencing of the complete cDNA confirmed reported that the MTUS1 gene contains 17 coding exons which its identity with Genbank accession number NM_001001924. The can be alternatively spliced to generate three major transcripts full-length ATIP3 sequence was then PCR-amplified and designated ATIP1, ATIP3 and ATIP4 [19,20]. By real-time RT- subcloned into the BamHI and XhoI restriction sites of the PCR, we showed that ATIP1 and ATIP3 are widely distributed in pEGFP-C1 vector (Clontech) so that the initiating codon of ATIP3 normal human tissues, while expression of ATIP4 is restricted to is in frame with the carboxy-terminal end of the Green Fluorescent the central nervous system. High evolutionary conservation of protein (GFP). ATIP amino-acid sequences suggests a pivotal role for these Transfection of plasmid cDNA into HeLa, HeLa-H2B, MDA- proteins in cellular homeostasis [20], but to date only ATIP1 has MB-231 or MCF7 cells was performed for 24 h using Lipofecta- been cloned and characterized. This isoform has been identified as mine Reagent 2000 (Invitrogen) as described by the manufacturer. an interacting partner of the angiotensin II AT2 receptor involved in trans-inactivation of the EGF receptor and subsequent Small interfering RNA (siRNA) was transfected at 50 nM for 72 h inhibition of extracellular regulated ERK kinase activity and cell using Lipofectamine reagent 2000. Control siRNA duplex (non # # proliferation [13,19,21]. However, the expression and function of targeting pool 1), and specific MTUS1 siRNA 1 (on-target plus ATIP proteins in breast cancer have not yet been explored. smart pool, NM020749) and siRNA#2 (on-target plus siRNA The present study investigates the expression levels and duplex, sens strand : UGG CAG AGG UUU AAG GUU A) were biological effects of MTUS1 gene products in infiltrating breast purchased from Dharmacon (Chicago, IL, USA). Silencing cancer. We show that ATIP3 is the major MTUS1 splice variant efficiency was measured by western blot and real time RT-PCR. whose expression is significantly reduced in invasive breast tumors of high histological grade and triple-negative phenotype. Molec- Real-time RT-PCR ular cloning and functional studies presented here reveal that Total RNA was extracted from tumor samples by the cesium ATIP3 is a novel mitotic spindle-associated protein that reduces chloride protocol [28], and from cancer cell lines using Trizol breast cancer cell division in vitro and in vivo. reagent (Invitrogen) according to the manufacturer’s instructions. First-strand cDNAs were synthesized using oligo(dT) primer with TM Methods Superscript reverse transcriptase (Invitrogen) for 60 min at 37uC. cDNA (5 ng) was used for real-time PCR analysis using Fast Breast tumor samples Start SYBR Green I reagent (Roche) and LightCycler instrument. Infiltrating ductal primary breast carcinomas were obtained Each reaction was carried out in duplicate. EEF1G and PP1A from patients included between in the prospective database cDNAs served as internal controls. Gene-specific primer pairs initiated in 1981 by the Institut Curie Breast Cancer group. All located in different exons (supplemental Table S1) were designed patients gave original verbal consent on the use of tumor with Oligo6 software. Results are expressed relative to EEF1G specimens for research purposes. Eleven normal breast specimens values.

PLoS ONE | www.plosone.org 2 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Immunohistochemistry Tumors were fixed in 4% formaldehyde solution and paraffin- Immunohistochemistry was performed on 5 mm sections from embedded for histologic analysis. formalin-fixed, paraffin-embedded tissue of the breast cancer samples. Heat-mediated antigen retrieval was performed in EDTA Immunofluorescence microscopy and live cell imaging buffer pH 9 in water bath for 30 min. Immunostaining was Exponentially growing cells plated on coverslips were fixed in performed on a Dako autostainer using a peroxidase-labeled ice-cold methanol for 5 min prior to incubation for 1 hr at room polymer-based detection system (Envision plus, Dako) and temperature with primary antibodies. SK-MES cells were diaminobenzidine as a chromogen. Monoclonal anti-MTUS1 incubated with human anti-alpha-tubulin antibodies clone F2C antibodies (Abnova) were diluted 1:40 and incubated overnight at diluted 1:10 [29] and mouse anti-MTUS1 monoclonal antibodies 4uC. Slides were counterstained with hematoxylin. MCF7 cells diluted 1:10 (Abnova, clone 1C7). HeLa, MDA-MB-231 or MCF7 transfected with GFP-ATIP3 were used as a positive control and cells transiently transfected (24 hrs) with pEGFP-C1 or GFP- non-transfected MCF7 cells were considered as negative control. ATIP3 were incubated with mouse anti-alpha-tubulin monoclonal MTUS1 immunoreactivity in tissue sections was scored based on antibodies (Sigma) diluted 1:500. After extensive washing, cells the percentage of positive cells and the intensity (1 to 3) of the were incubated for 30 min at room temperature with appropriate staining. Tumors were considered MTUS1-positive when more secondary Cy-2-conjugated anti-human and/or Cy-3-conjugated than 60% of the cells showed intense (2–3) staining, and were anti-mouse antibodies (Jackson Laboratories, Interchim, France) considered MTUS1-negative when less than 30% of the cells diluted 1:500. Coverslips were mounted on glass slides using showed weak (0–1) immunostaining. Glycergel Mounting Medium (DakoCytomation) and examined For analysis of xenografts, 3 mm sections were cut from with a Leica TCS SP2 AOBS confocal microscope using formalin-fixed, paraffin-embedded tissue blocks, then counter- appropriate filters. Image analysis was performed with NIH stained with hematoxylin-eosin and examined under an inverted ImageJ software. microscope. For live cell imaging, images were acquired on a spinning disk microscope (one image taken every 5 min for 36 hrs following Cell proliferation assays transient transfection of HeLa-mCherryH2B cells) as described Cells were seeded in quadruplicate in 96-well plates at the [30]. Multi-dimensional acquisitions were performed in time-lapse density of 5.103 cells per well. Cell proliferation was analyzed by mode using Metamorph 7.1.7 software. incubation for 4 h at 37uC with 1 mg/ml tetrazolium salt MTT (Sigma). Cleavage products (formazan) were solubilized with Microtubules binding in cell extract DMSO and optical density was measured at 560 nm, as Microtubule cosedimentation assay was performed as described recommended by the manufacturer. DNA synthesis was measured [31] with modifications. Cells were incubated for 20 min at 4uCin using a colorimetric immunoassay (Roche) after incorporation of PEM buffer (100 mM PIPES, pH 6.9, 1 mM MgSO4, 1 mM 5-bromodeoxyuridine (BrdU) for 4 h at 37uC. Incorporated BrdU EGTA), scraped and sonicated prior to centrifugation at was quantified by optical density reading at 405 nm according to 15000 rpm for 10 min, 4uC. Clarified samples were incubated manufacturer’s instructions. with taxol (20 mM) in the presence of GTP (1 mM) and DTT For clonogenicity experiments, MCF7 cells were transfected for (1 mM) for 45 min at 37uC and were spun at 70 000 g for 30 min 24 h with 0.2 mg pEGFP-C1 vector or 2 mg of GFP-ATIP3 cDNA. at 30uC through a cushion buffer containing 40% glycerol, 20 mM Similar transfection efficiency (50%) was assessed by FACS taxol and 1 mM GTP. The supernatant (S) and pellet (P) fractions analysis and cell viability was verified using trypan blue. Cells were collected separately and subsequently immunoblotted with were plated at various dilutions in 6 well plates and transfectants anti-GFP (Sigma) or anti-MTUS1 (Abnova) antibodies. Blots were were selected by adding geneticin (G418, 1 mg/ml, Gibco) in fresh reprobed using anti-alpha-tubulin (Sigma) antibodies. medium twice a week for 3 weeks. Resistant colonies were stained with 0.5% cristal violet (Sigma) and counted. Statistical Analysis For anchorage-independent growth assays, stable transfectants In all studies, values are expressed as mean 6 standard deviation were seeded onto 6-well plates on 0.4% agar gel in appropriate (SD). Statistical analyses were performed by unpaired Student’s t medium supplemented with 20% FCS, over a bottom layer of test and Tukey-Kramer’s test using JMP7 software. Differences 0.6% low-melting temperature agar gel. Cells were grown for 4 were considered statistically significant at p,0.05. weeks and the formed colonies were counted under an inverted microscope. Results In vivo tumorigenicity assay MTUS1 gene expression is reduced in invasive breast Five weeks old female immunodeficient CB17/scid mice were cancer samples purchased from Harlan (UK). A slow-release 17-beta-estradiol The expression of MTUS1 transcripts was analyzed in a series of pellet (60-day release, 0.72 mg per pellet, Innovative Research of 151 infiltrating ductal breast carcinomas and eleven normal breast America, Florida, USA) was implanted subcutaneously into the tissues. The intensities of each of three specific MTUS1 probesets intrascapular region of each mouse three days before breast cancer (212096_s_at; 212093_s_at; 239576_at) were obtained from our cell inoculation. MCF7 cell clones (56106 cells in 100 ml PBS) previous U133A Affymetrix DNA array study [22] and compared were injected subcutaneously into both flanks. All procedures were to histologic characteristics of the tumors and clinical data for the performed in accordance with institutional guidelines established patients (supplemental Table S2). As shown in Fig. 1A and Table 1, by the Ministe`re de l’Agriculture et de la Foreˆt, Direction des MTUS1 expression levels were significantly reduced in breast services ve´te´rinaires (Paris, France). Tumor diameter was cancer samples as compared to normal tissue. MTUS1 levels were measured twice weekly using a calipper along two orthogonal significantly lower in high grade tumors (grade III) as compared to axes: length (L) and width (W). The volume (V) of tumors was grades I (p,0.0001) and II (p = 0.0032), but did not differ estimated by the formula V = L6(W2)/2. Mice were sacrificed on significantly between tumors of grades I and II (p = 0.06). Similar day 38 and tumors were excised, weighed and photographed. results were obtained with all three MTUS1 probesets (Fig. 1A and

PLoS ONE | www.plosone.org 3 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer supplemental Fig.S1). As presented in Table 1, a two- to ten-fold molecular breast tumors classified clinically according to immu- reduction in MTUS1 expression was observed in 47.7% of the nohistochemical detection of the surrogate markers ER, PR and tumors, and the number of samples underexpressing MTUS1 HER2. As shown in Fig. 1C, MTUS1 Affymetrix probeset increased with the histological grade : 38.8%, 47.5% and 84.6% in intensities were significantly lower in a majority (83.3%) of triple grade I, II and III, respectively. Low levels of MTUS1 also negative (ER2/PR2/HER22) carcinomas as compared to significantly correlated with tumors that developed metastasis at luminal (ER+) and HER2+ tumor subgroups (Table 1), which distant sites, but not with the occurrence of axillary lymph node supports the conclusion that reduced levels of MTUS1 are metastasis (Fig. 1B). MTUS1 levels were then compared among associated with breast cancer aggressive subtypes.

Figure 1. ATIP3 down-regulation in invasive breast carcinomas. A–B. Comparison of MTUS1 (U133A Affymetrix 212096_s_at) probeset data intensities in (A) normal breast tissue and 151 invasive breast tumors classified according to histological grade (I, II, III), (B) breast tumors classified according to the occurrence of distant metastasis or axillary lymph node (ALN) metastasis and (C) molecular subgroups of breast tumors defined by immunodetection of surrogate markers ER/PR+ (luminal), HER2+ (HER22like) and ER2,PR2, HER22 (triple negative, TN). Probeset intensities were calculated using Affymetrix Raw MAS5.0 default settings. The number of samples is indicated below in brackets. *p,0.001; **p,0.0001 compared to normal tissue; ap,0.0001 compared to grade I; bp,0.005 compared to grade II; cp,0.0001 compared to ER+/PR+; dp,0.05 compared to HER2+. D. Immunohistochemistry on a breast tumor section of histological grade I and adjacent normal breast tissue (upper panel), and two representative grade III breast cancer sections (lower panel) using anti-MTUS1 monoclonal antibodies. A bar represents 100 mm. E–G. Correlation between MTUS1 (212096_s_at) probeset intensities and real-time RT-PCR expressed relative to internal control EEF1G in 29 representative breast tumor samples. Oligonucleotides were designed to amplify total ATIP transcripts (‘‘MTUS1’’) (E), ATIP3 transcripts (F), or ATIP1 transcripts (G). doi:10.1371/journal.pone.0007239.g001

PLoS ONE | www.plosone.org 4 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Table 1. MTUS1 expression levels in invasive breast carcinomas.

Number of tumors Number of tumors Raw Mas5 Median Raw Mas5 [range] Raw Mas5 Mean+/2SD % underexpressed underexpressed

Normal 11 2508 [1842.9–3388.6] 2561.5+/2468.8 Grade I 64 1563 [532.2–3414.9] 1646.7+/2651.5 38.8 21 Grade II 61 1354 [636.4–3105.5] 1438+/2595.6 47.5 29 Grade III 26 912 [321.5–3030.7] 1005.3+/2616.3 84.6 22 ER+/PR+ 112 1443 [616.8–3414.9] 1539.3+/2640.5 42.8 48 HER2+ 18 1198.7 [599.6–3030.7 1320.3+/2597.4 55.5 10 TN 12 818.5 [321.5–1635.4] 818.5+/2416.4 83.3 10 All tumors 151 1353.7 [321.5–3414.9] 1452+/2659 47.7 72

Raw Mas5 Affymetrix values (probeset 212096_s_at) in cancer samples classified according to their histological grade (I, II, III) and molecular subtypes ER+/PR+, HER2+, TN (triple negative). Tumors with Affymetrix values being more than two-fold lower than those in normal samples were considered underexpressed. doi:10.1371/journal.pone.0007239.t001

Immunohistochemical analyses were undertaken to examine the 474 and MDA-MB-468 expressed moderate levels of MTUS1 cellular distribution and expression of MTUS1 at the protein level in transcripts. Highest levels of MTUS1 expression were found in the situ. Immunohistochemistry performed on normal breast tissue lung cancer cell line SK-MES, that was therefore selected for further sections using monoclonal anti-MTUS1 antibodies revealed a strong study. As for primary tumors, ATIP3 expression in cancer cells cytosolic staining in luminal epithelial cells of the breast (Fig. 1D). The paralleled that of total MTUS1 mRNA, while ATIP1 remained same results were obtained using polyclonal anti-ATIP antibodies undetectable in every cell line examined. In agreement with our RT- described previously [19] (data not shown). MTUS1 immunostaining PCR studies, western blot experiments using anti-MTUS1 antibodies was then analyzed in 20 breast cancer samples for which paraffin- revealed a polypeptide of 170 KDa corresponding to ATIP3 in embedded tissue sections were available (Fig. 1D). Selected tumors cancer cell lines MDA-MB-468 and SK-MES, but not in MCF7 nor were representative of all histological grades (10, 3 and 7 tumors of MDA-MB-231 (supplemental Fig.S2B). grade I, II and III, respectively), and main molecular subtypes (12 The functional consequence of ATIP3 underexpression was ER+,3HER2+ and 3 triple negative tumors). MTUS1 immuno- investigated by transfection of specific siRNA into the ATIP3- staining in tumor sections was scored and compared to MTUS1 positive MDA-MB-468 cell line. Two different sequences of mRNA expression levels according to Affymetrix probeset intensities. siRNA successfully silenced ATIP3 expression in MDA-M468 cells As presented in supplemental Table S3, tumors showing low levels of both at the mRNA and protein level (Fig. 2A, left panels). As MTUS1 transcripts by DNA array and/or real-time RT-PCR shown in Fig. 2A (right panel), ATIP3 knock-down by both analysis also displayed undetectable or weak MTUS1 staining, siRNAs led to a significant increase (66.465.3% and 63610.2% whereas those expressing high levels of MTUS1 mRNA were upon transfection of siRNA#1 and siRNA#2, respectively) in attributed a high score by immunohistochemistry. Thus, results on MDA-MB-468 breast cancer cell proliferation, pointing to an anti- protein expression paralleled gene array analysis. proliferative effect of ATIP3. Results obtained by DNA microarray were further confirmed and To further confirm the inhibitory effects of ATIP3 on breast cancer extended by real-time RT-PCR. We selected a panel of 29 cell proliferation, a GFP-ATIP3 fusion protein was generated and representative tumor samples whose MTUS1 Affymetrix probesets expressed in ATIP3-negative cells MDA-MB-231 and MCF7. intensities ranged from low (n = 11), moderate (n = 13) to high (n = 5). Independent stable transfectants (clone 3A1 in MDA-MB-231, and Expression levels of total MTUS1 transcripts, measured by quantitative clones HC1, HC6, HC7 in MCF7) expressed GFP-ATIP3 at levels PCR using oligonucleotides located in the 39 exons shared by all ATIP similar to those of endogenous ATIP3 in MDA-MB-468 (Fig. 2B and members, were found to significantly correlate (r2 = 0.77) with 2D, left panel). As shown in Fig. 2B (right panels), cell proliferation Affymetrix probeset intensities (Fig. 1E). In order to discriminate measured by BrdU incorporation and MTT assay was reduced by between different MTUS1 spliced variants, real-time RT-PCR was 2862.3% and 3065%, respectively, in GFP-ATIP3-transfected clone designed to specifically amplify ATIP1 and ATIP3 mRNAs using 3A1 as compared to non transfected and GFP-expressing MDA-MB- oligonucleotides located in their specific 59 exons, respectively. Since 231 cells. Similar results were obtained using stably transfected MCF7 ATIP4 transcripts are exclusively expressed in the central nervous clones HC1, HC6 and HC7 (data not shown). system [20], these were not included in the present study. Significant Clonogenicity experiments were then conducted to evaluate the correlation was observed between Affymetrix probeset intensities and consequences of ATIP3 expression on cell-cell contact growth expression levels of ATIP3 (r2 = 0.83) (Fig. 1F), but not ATIP1 inhibition. As shown in Fig. 2C, the number of colonies resistant to transcripts (r2 = 0.01) (Fig. 1G), indicating that ATIP3 is the major neomycin were reduced by 85% upon transfection of MCF7 cells MTUS1 splice variant whose expression is regulated in breast cancer. with GFP-ATIP3 as compared to GFP. To investigate the effects of ATIP3 on anchorage-independent ATIP3 inhibits breast cancer cell proliferation growth, which is a hallmark of tumorigenic properties in vitro, As a pre-requisite to functional studies, a panel of human cancer stably transfected breast cancer cells were allowed to grow for four cell lines was analyzed for MTUS1 expression by real-time RT-PCR weeks in soft agar. As shown in Fig. 2D, the number of colonies (Supplemental Fig. S2A). Nine out of 13 breast cancer cell lines grown in soft agar were reduced by 75% in MCF7 cell clones including MDA-MB-231, MCF7, Hs578T, BT549, MDA-MB-453, transfected with GFP-ATIP3 (HC1 and HC6) as compared to T-47D, ZR-75.1, MDA-MB-361 and SK-BR3 showed low to GFP. Similar results were obtained using stably transfected cell undetectable levels of MTUS1,whereasCAMA-1,ZR-75.30,BT- clones HC7 and 3A1 (data not shown).

PLoS ONE | www.plosone.org 5 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Figure 2. ATIP3 inhibits cancer cell proliferation. A. ATIP3 silencing in MDA-MB-468 cells transfected with control siRNA or specific ATIP3 siRNA#1 or siRNA#2 for 72 hours. Left panel: real-time RT-PCR using ATIP3-specific primers relative to EEF1G, normalized to non transfected cells (NT). Middle panel: immunoblotting with anti-MTUS1 antibodies showing ATIP3 at 170 kDa, and reprobing with anti-alpha-tubulin antibodies for internal control. Right panel: MTT assay. Results are expressed as percent of MTT incorporation in control siRNA-transfected cells at time 96 hours. Shown are the results of two representative experiments out of four performed in quadruplicate. **p,0.0001. B. Cell proliferation in stable MDA-MB- 231 transfectants. Left panel : immunoblotting of total cell lysates from MDA-MB-231 cells non transfected (NT) or stably expressing GFP or GFP-ATIP3 (clone 3A1) at levels comparable to those in MDA-MB-468. Blots were probed with anti-MTUS1 (upper panel), anti-GFP (middle panel), anti-alpha- tubulin antibodies (lower panel). Arrows indicate the migration of GFP-ATIP3 (200 kDa), ATIP3 (170 kDa), GFP (25 kDa), alpha-tubulin (50 kDa). Middle and right panels: measurement of BrdU incorporation and MTT assay. Shown is one representative experiment out of three performed in quadruplicate. **p,0.0001; *p,0.001. C. Colony formation of GFP- or GFP-ATIP3-transfected MCF7 cells plated at different densities. Shown is one representative experiment out of five performed in duplicate. Right panel : quantification of the number of colonies per well. **p,0.0001. D. Immunoblotting of total lysates from MCF7 cells left untransfected (NT) or stably transfected with GFP or GFP-ATIP3 (clones HC1 and HC6), revealed as in (B). Middle panel : MCF7 stable transfectants grown in soft agar for four weeks. Shown are photomicrographs of one representative experiment out of three performed in triplicate. A bar represents 50 mm (upper panel) and 10 mm (lower panel). Right panel : quantification of the number of colonies per field counted under an inverted microscope. **p,0.0001. doi:10.1371/journal.pone.0007239.g002

ATIP3 expression reduces tumor growth in vivo mice injected with HC1, HC6 and HC7 as compared to GFP- Stably transfected MCF7 clones (HC1, HC6 and HC7, or GFP) expressing MCF7 cells. Histological examination of the tumor were injected subcutaneously into both flanks of immunodeficient nodules excised on day 38 confirmed the presence of tumor cells in mice, and tumor growth was monitored twice a week. At day 30, each case (Fig. 3E). However a two-fold reduction in the number of only 17% (6/36) of tumor xenografts developed in mice injected cancer cells undergoing active division, assessed by microscopic with GFP-ATIP3-transfected clones (HC1, HC6 or HC7), as examination of mitotic cells (Fig. 3F), was observed in tumors compared with 82% (18/22) in mice injected with GFP-transfected expressing GFP-ATIP3 as compared to GFP alone. Thus, restoring MCF7 cells (Fig. 3A). The time-course of tumor progression (Fig. 3B) endogenous levels of ATIP3 expression leads to reduced breast and size of the tumors (Fig. 3C, 3D) were significantly reduced in cancer cell proliferation and delayed tumor growth in vivo.

PLoS ONE | www.plosone.org 6 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Figure 3. ATIP3 suppresses tumor growth in vivo.A.Number of tumor xenografts (in percent) developing after s.c. injection of GFP-ATIP3 expressing MCF7 clones (HC1, HC6, HC7) or GFP-transfected MCF7 cells (stable clone or pool) into immunodeficient mice. B. Time-course of tumor progression in mice injected with GFP or GFP-ATIP3 cell clones as in A. Tumor size was measured twice weekly as indicated in the methods. C. Mean weight (in grams) of tumors excised on day 38 after injection of MCF7 cells expressing GFP (clone and pool, n = 18) and GFP-ATIP3 (clones HC1, HC6, HC7, n = 11). D. Pictures of representative tumors excised on day 38 after injection of MCF7 cell clones as in C. E. Representative histologic analysis of tumors grown from GFP and GFP-ATIP3 expressing MCF7 cells, after hematoxylin-eosin staining. A bar represents 100 mm. F. Representative photomicrographs of cancer cells as in E, stained by hematoxylin-eosin (magnification663). Arrows show cancer cells undergoing active mitosis. A bar represents 20 mm. Histogram shown on the right is a quantification of the number of mitotic cells per field counted under an inverted microscope. The number of fields analyzed is indicated below in brackets. **p,0.0001. doi:10.1371/journal.pone.0007239.g003

ATIP3 associates with microtubules at the mitotic spindle during mitosis (not shown), and stains the To further characterize ATIP3 at the molecular level, its intercellular bridge during cytokinesis (Fig. 4B). subcellular localization was analyzed in SK-MES cancer cells that Microtubule co-sedimentation assays were then performed on express high levels of ATIP3 and no detectable ATIP1. Confocal cells transfected with either GFP-ATIP3 or GFP. In this assay, microscopy analyses (Fig. 4A) allowed the detection of endogenous microtubules are stabilized in the presence of taxol and GTP, and ATIP3 proteins as filamentous structures in the cytoplasm, with a proteins that associate with microtubules are captured in the pellet staining pattern similar to that of microtubules. Incubation with fraction. As shown in Fig. 4C (upper panel), the whole pool of GFP- anti-alpha-tubulin antibodies revealed colocalization of ATIP3 ATIP3 co-sediments with stabilized microtubules in the pellet with the microtubule network during interphase. During mitosis, whereas the soluble GFP protein is only detected in the supernatant endogenous ATIP3 staining decorates the mitotic spindle in fraction. Co-sedimentation of GFP-ATIP3 with microtubules was metaphase and anaphase, and remains strongly associated with the observed in both MCF7 and MDA-MB-231 cells either stably or intercellular bridge during cytokinesis (Fig. 4A). To confirm that transiently transfected with GFP-ATIP3 (data not shown). the ATIP3 isoform indeed colocalizes with microtubules, human Whether endogenous ATIP3 also co-sediments with microtu- cell lines were transiently transfected with the GFP-ATIP3 fusion bules was further evaluated in SK-MES, MDA-MB-468 and protein and analyzed by fluorescence microscopy. As shown in CAMA-1 cancer cells, that express high to moderate levels of Fig. 4B, GFP-ATIP3 colocalizes with alpha-tubulin at the ATIP3 (supplemental Fig.S2). As shown in Fig. 4C (lower panel), the microtubule cytoskeleton and microtubule organizing center whole pool of endogenous ATIP3 was detected in the microtubule (MTOC) in cells expressing low levels of the fusion protein. pellet and no residual soluble protein was found in the supernatant. Higher levels of expression of the GFP-ATIP3 construct led to the Thus, ATIP3 associates with stabilized microtubules. formation of bundles which is a characteristic feature of stabilized microtubules, suggesting that ATIP3 may regulate the dynamics of ATIP3 promotes prolonged mitosis microtubule polymerization. The same intracellular localization The localization of ATIP3 at the mitotic spindle and was observed upon transfection of GFP-ATIP3 into HeLa, RPE1, intercellular bridge, together with its anti-proliferative effects in MDA-MB-231 and MCF7 cells (not shown). As observed for vitro and in vivo, prompted us to investigate the effects of this protein endogenous ATIP3, the GFP-ATIP3 fusion protein also localizes in essential steps of cell division. To this end, GFP-ATIP3 was

PLoS ONE | www.plosone.org 7 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Figure 4. ATIP3 associates with microtubules. A–B. Representative photomicrographs of confocal microscopy analysis showing the cellular distribution of (A) endogenous ATIP proteins in SK-MES cancer cells and (B) transiently transfected (24 h) GFP-ATIP3 fusion protein, stained with anti- MTUS1 (green) and anti-alpha-tubulin (red) antibodies. A bar represents 10 mm. C. Microtubule co-sedimentation assay performed on MDA-MB-231 stable transfectants expressing GFP-ATIP3 or GFP (upper panel) and SK-MES, MDA-MB-468 and CAMA-1 tumor cells expressing endogenous ATIP3 (lower panel). Immunoblots were revealed using anti-GFP or anti-MTUS1 antibodies, and reprobed with anti-alpha-tubulin antibodies. Molecular weights are indicated on the left. L: total cell lysate; S: supernatant; P: pellet. doi:10.1371/journal.pone.0007239.g004 transiently transfected into HeLa-H2B cells, that stably express a metastasis, indicating a close relationship between reduced ATIP3 red fluorescent protein-tagged Histone H2B and constitute a useful expression and breast cancer aggressiveness. In addition, in tumors tool to monitor mitosis by live cell imaging. Time-lapse showing a triple-negative phenotype (ER2 PR2 HER22) ATIP3 fluorescence videomicroscopy analyses revealed the association levels are significantly lower than those measured in luminal (ER+) of GFP-ATIP3 with duplicated centrosomes during prophase, and and HER2+, suggesting that ATIP3 may represent a novel further confirmed its localization at the mitotic spindle and molecular marker for poor outcome of breast cancer. intercellular bridge at different steps of mitosis in living cells Immunohistochemical analyses revealed a cytosolic ATIP (Fig. 5A, supplemental video S1 and supplemental video S2). In localization in luminal epithelial cells of normal breast tissue. addition, measuring the time-lapse between chromosome conden- ATIP protein immunostaining intensity paralleled its mRNA levels sation and the end of cytokinesis revealed that cells expressing determined both by gene microarray and real-time RT-PCR. GFP-ATIP3 need a longer time to achieve mitosis (1876138 min, Thus, immunodetection of ATIP proteins in breast cancer samples n = 35) as compared to non-transfected cells (81612 min, n = 9) may be a simple and useful tool for pathologists to distinguish and cells expressing GFP (5967 min, n = 16) (Fig. 5B). More between ATIP-positive and negative tumors. This may be of specifically, the time spent in metaphase (from chromosome particular interest for the identification of a subset of triple condensation in prometaphase to the beginning of anaphase) was negative breast tumors having lost the expression of ATIP3. significantly increased in GFP-ATIP3 (135.96129.7 min, n = 33) Indeed, in contrast to the ER+ and HER2+ tumors subtypes that as compared to GFP-transfected cells (1965.6 min, n = 16) can be efficiently treated with hormonal and anti-HER2 therapy, (Fig. 5C). Thus, ATIP3 overexpression delays the progression of respectively, triple negative tumors lack targeted treatments. We mitosis by prolonging the step of metaphase. show here that restoring ATIP3 expression in ATIP3-negative breast cancer cell lines leads to reduced cancer cell proliferation, Discussion clonogenicity and anchorage-independent growth. Furthermore, ATIP3 re-expression lowers the incidence, time-course and size of The results presented here identify ATIP3 as a novel anti- tumor progression in xenograft models in vivo, suggesting that mitotic protein whose expression is reduced in infiltrating breast ATIP3 may represent an interesting candidate for future targeted cancer. Low levels of expression of ATIP3 are associated with high therapies of triple negative breast tumors. Further studies will be histological grade of the tumor and the occurrence of distant necessary to investigate the feasibility of in vivo methods aimed at

PLoS ONE | www.plosone.org 8 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Figure 5. ATIP3 delays mitosis. A. Representative photomicrographs from live cell imaging of HeLa cells stably expressing mCherry-histone 2B (in red) and transiently transfected with GFP-ATIP3 (in green). Time (in hrs) is indicated above or below each panel. B. Time (in min) necessary to achieve mitosis in control, GFP-transfected and GFP-ATIP3 transfected cells. The number of cells analyzed is indicated below in brackets. **p,0.0001. C. Time (in min) between chromosome condensation and the onset of anaphase in GFP-transfected and GFP-ATIP3 transfected cells. The number of cells analyzed is indicated below in brackets. **p,0.0001. doi:10.1371/journal.pone.0007239.g005 re-expressing ATIP3, or active truncated domains of the protein, to the identification of potential driver mutations that may lead to by using tumor-targeted systemic delivery of nanocomplexes as a loss of function of the ATIP3 protein in cancer. recently reported for the tumor suppressor RB94 [32]. Our results show that ATIP3 is a novel microtubule-associated ATIP3 belongs to a family of proteins (ATIP1 to ATIP4) protein localized at the centrosome, mitotic spindle and intercel- encoded by alternative splicing of MTUS1, a candidate 8p22 lular bridge, and that its overexpression delays the progression of tumor suppressor gene [20]. Previous studies have reported that mitosis in living cells. ATIP3 thus appears as a new member of a MTUS1 expression levels are reduced in cancers from pancreas, functional family of microtubule-associated proteins, including colon, ovary and head-and-neck [13–16]. However, the question adenomatous polyposis coli APC [34], LZTS1 [35] and RASSF1A of which MTUS1 transcript is down-regulated had not been [31,36], that interfere with the microtubule cytoskeleton to investigated, nor had the in vivo consequence of ATIP re- regulate essential steps of mitosis and have been clearly identified expression been analyzed until now. We report here for the first as tumor suppressors down-regulated or mutated in breast cancer. time that ATIP3 is the major MTUS1 isoform whose expression is The possibility that ATIP3 may be involved in cell division via a altered in invasive breast cancer. Our results showing that ATIP3 regulation of spindle dynamics is consistent with its localization at re-expression functionally reduces breast cancer cell growth in vitro centrosomes, spindle poles and spindle microtubules, and is and in vivo may argue in favor of a tumor suppressor effect of supported by its effects on microtubule stability (our unpublished ATIP3. However, whether MTUS1 corresponds to a ‘‘bona fide’’ observations). ATIP3 may thus represent a novel interesting target tumor suppressor gene inactivated by epigenetic or genomic for rational anti-mitotic therapies based on proteins associated alterations remains to be established. Of interest, our previous with the mitotic spindle [37,38]. Furthermore, the observation that mutational analysis of MTUS1 in a series of hepatocellular ATIP3 promotes prolonged mitosis by extending the time of carcinomas and cell lines has identified five potential gene metaphase potentially reflects modulation of spindle checkpoint mutations which are all located in exons specific to the ATIP3 signalling and provides a molecular basis for elucidating the transcript [33]. We and others have also identified MTUS1 gene intracellular functions of ATIP3 in cell division. Thus, the novel variations located on ATIP3-specific exons in head-and-neck microtubule associated protein ATIP3 identified here may carcinoma cell lines [33,16]. Furthermore, Frank et al. have represent a useful prognostic biomarker of agressive breast cancer recently reported significant association between copy number and a promising candidate for the development of new targeted variants of MTUS1 exon 4 - specific to the ATIP3 transcript - and therapeutic strategies. familial breast cancer risk [18], further suggesting that alterations of the ATIP3 isoform may be associated with carcinogenesis. To Supporting Information note, the functional consequences of MTUS1 gene alterations could not be examined until now due to a lack of knowledge on Figure S1 MTUS1 down-regulation in invasive breast carcino- ATIP3 intracellular actions. By providing a molecular and mas. U133A Affymetrix MTUS1 probesets (239576_at; functional analysis of ATIP3, the present study brings a major 212093_s_at) intensities in normal breast tissue and 151 invasive contribution to the characterization of MTUS1 gene products and breast tumors classified according to histological grade (I, II, III).

PLoS ONE | www.plosone.org 9 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

Probeset intensities were calculated using Affymetrix Raw MAS5.0 Found at: doi:10.1371/journal.pone.0007239.s005 (0.01 MB default settings. The number of samples is indicated below under XLS) brackets. Video S1 Live cell imaging of GFP-ATIP3 in HeLa-H2B cells. Found at: doi:10.1371/journal.pone.0007239.s001 (1.57 MB TIF) HeLa-H2B cells stably expressing mcherry-Histone 2B were Figure S2 MTUS1/ATIP expression in human cancer cell lines. transfected for 24 hrs with GFP-ATIP3 and imaged by spinning A. Real-time RT-PCR on 14 human tumor cell lines using disk confocal microscopy in a single focal plane every five minutes MTUS1, ATIP1 or ATIP3 primers as defined in the methods, for 36 hrs following transfection. GFP-ATIP3 fluorescence expressed relative to internal control EEF1G. B. Immunoblotting labeling (in green) associates with the centrosomes, microtubule of total cell lysates from cell lines MCF7, MDA-MB-231, MDA- spindle, cleavage furrow and intercellular bridge during mitotic MB-468 and SKMES using anti-MTUS1 monoclonal antibodies. progression of live transfected cells. Fluorescence labeling of DNA Blots were reprobed with anti-ezrin for internal control. Arrows on is in red. the left indicate apparent molecular weights of endogenous ATIP3 Found at: doi:10.1371/journal.pone.0007239.s006 (3.83 MB (170 kDa) and ezrin (80 kDa). MOV) Found at: doi:10.1371/journal.pone.0007239.s002 (1.46 MB TIF) Video S2 Live cell imaging of GFP-ATIP3 in HeLa-H2B cells. Table S1 Sequences and gene location of oligonucleotides used HeLa-H2B cells stably expressing mcherry-Histone 2B were in real-time RT-PCR. transfected for 24 hrs with GFP-ATIP3 and imaged by spinning Found at: doi:10.1371/journal.pone.0007239.s003 (0.03 MB disk confocal microscopy in a single focal plane every five minutes DOC) for 36 hrs following transfection. GFP-ATIP3 fluorescence Table S2 MTUS1 affymetrix probesets intensities in 151 labeling (in green) associates with the centrosomes, microtubule invasive breast carcinomas. Intensities of three MTUS1 probesets spindle, cleavage furrow and intercellular bridge during mitotic (212093_s_at; 212096_s_at; 239576_at) in 151 infiltrating ductal progression of live transfected cells. Fluorescence labeling of DNA carcinomas (IDC) and 11 normal breast tissues. Values (Raw is in red. MAS05 setting defaults) are compared to the histological grade Found at: doi:10.1371/journal.pone.0007239.s007 (2.20 MB (Easton and Ellis, EE) of the tumors and the occurrence of axillary MOV) lymph node (ALN) and distant metastasis. Immunohistochemical detection of surrogate markers : epidermal growth factor receptor Acknowledgments 1 (EGFR), human epidermal growth factor receptor 2 (HER2), estrogen receptor (ER), and progesterone receptor (PR), is as We wish to thank Dr. Nicolas Stransky (Institut Curie, Paris) for sharing described in the Methods. EGFR and HER2 values were scored 0 Affymetrix probeset data analysis, and Dr. Vale´rie Doye (Institut Jacques to 100 according to the percentage of stained cells. ND : not Monod, Paris) for the generous gift of HeLa cells expressing mCherry- histone. We thank the Institut Curie breast cancer group (head: Brigitte determined. Sigal-Zafrani) : Bernard Asselain, Marc Bollet, Franc¸ois Campana, Paul Found at: doi:10.1371/journal.pone.0007239.s004 (0.05 MB Cottu, Patricia de Cre´moux, Ve´ronique Die´ras, Alain Fourquet, Youlia XLS) Kirova, Jean-Yves Pierga, Anne Vincent-Salomon, Re´my Salmon, Table S3 Comparison of MTUS1 mRNA levels and protein Dominique Stoppa-Lyonnet, Anne Tardivon, Fabienne Thibault, and Fabien Valet, for helpful discussion. We greatly acknowledge Isabelle immunostaining in 20 invasive breast carcinomas. Affymetrix Loı¨odice, Gae¨l Le Gouevec, Franck Letourneur (Functional genomics 212096_s_at probeset intensities (Raw MAS5.0) values, total platform, Institut Cochin, Paris) and Pierre Bourdoncle (Imaging platform, MTUS1 mRNA levels measured by real-time PCR relative to Institut Cochin, Paris) for their invaluable technical help. EEF1G internal control, and immunohistochemical (IHC) results using anti-MTUS1 monoclonal antibodies. MTUS1 mmunostain- Author Contributions ing was scored as described in the Methods, according to the percentage (%) of positive tumor cells and intensity of the staining. Conceived and designed the experiments: SRF ADT AD SC VM FP AVS Values are compared to histological grade (I, II, III) of the tumor CN. Performed the experiments: SRF ADT AD SC NG HC AN NC VM FR BT. Analyzed the data: SRF ADT AD SC NG HC NC VM BT OD and immunodetection of surrogate markers (ER, PR, EGFR, FR FP AVS CN. Contributed reagents/materials/analysis tools: FR BSZ HER2) determined as described in the Methods. nd : not OD FR. Wrote the paper: SRF CN. determined.

References 1. Dowsett M, Dunbier AK (2008) Emerging biomarkers and new understanding different clinicopathological features and gene-expression subtypes of breast of traditional markers in personalized therapy for breast cancer. Clin Cancer cancer. Genes Chromosomes Cancer 45: 1033–1040. Res 14: 8019–8026. 9. Ade´laı¨de J, Finetti P, Bekhouche I, Repellini L, Geneix J, et al. (2007) Integrated 2. Rakha EA, Reis-Filho JS, Ellis IO (2008) Basal-like breast cancer: a critical profiling of basal and luminal breast cancers. Cancer Res 67: 11565–11575. review. J Clin Oncol 26: 2568–2581. 10. Hu X, Stern HM, Ge L, O’Brien C, Haydu L, et al. (2009) Genetic alterations 3. Sotiriou C, Pusztai L (2009) Gene-expression signatures in breast cancer. and oncogenic pathways associated with breast cancer subtypes. Mol Cancer N Engl J Med 360: 790–800. Res 7: 511–522. 4. Perou CM, Sørlie T, Eisen MB, van de Rijn M, Jeffrey SS, et al. (2000) 11. Chin L, Gray JW (2008) Translating insights from the cancer genome into Molecular portraits of human breast tumours. Nature 406: 747–752. clinical practice. Nature 452: 553–563. 5. Sørlie T, Perou CM, Tibshirani R, Aas T, Geisler S, et al. (2001) Gene 12. Hirano A, Emi M, Tsuneizumi M, Utada Y, Yoshimoto M, et al. (2001) Allelic expression patterns of breast carcinomas distinguish tumor subclasses with losses of loci at 3p25.1, 8p22, 13q12, 17p13.3, and 22q13 correlate with clinical implications. Proc Natl Acad Sci U S A 98: 10869–10874. postoperative recurrence in breast cancer. Clin Cancer Res 7: 876–882. 6. Parker JS, Mullins M, Cheang MC, Leung S, Voduc D, et al. (2009) Supervised risk 13. Seibold S, Rudroff C, Weber M, Galle J, Wanner C, et al. (2003) Identification predictor of breast cancer based on intrinsic subtypes. J Clin Oncol 27: 1160–1167. of a new tumor suppressor gene located at chromosome 8p21.3-22. FASEB J 17: 7. Chin K, DeVries S, Fridlyand J, Spellman PT, Roydasgupta R, et al. (2006) 1180–1182. Genomic and transcriptional aberrations linked to breast cancer pathophysiol- 14. Pils D, Horak P, Gleiss A, Sax C, Fabjani G, et al. (2005) Five genes from ogies. Cancer Cell 10: 529–541. chromosomal band 8p22 are significantly down-regulated in ovarian carcinoma: 8. Bergamaschi A, Kim YH, Wang P, Sørlie T, Hernandez-Boussard T, et al. N33 and EFA6R have a potential impact on overall survival. Cancer 104: (2006) Distinct patterns of DNA copy number alteration are associated with 2417–2429.

PLoS ONE | www.plosone.org 10 October 2009 | Volume 4 | Issue 10 | e7239 ATIP3 Protein in Breast Cancer

15. Lee S, Bang S, Song K, Lee I (2006) Differential expression in normal-adenoma- 26. Neve RM, Chin K, Fridlyand J, Yeh J, Baehner FL, et al. (2006) A collection of carcinoma sequence suggests complex molecular carcinogenesis in colon. Oncol breast cancer cell lines for the study of functionally distinct cancer subtypes. Rep 16: 747–754. Cancer Cell 10: 515–527. 16. Ye H, Pungpravat N, Huang BL, Muzio LL, Mariggio` MA, et al. (2007) 27. Kenny PA, Lee GY, Myers CA, Neve RM, Semeiks JR, et al. (2007) The Genomic assessments of the frequent loss of heterozygosity region on 8p21.3 morphologies of breast cancer cell lines in three-dimensional assays correlate approximately p22 in head and neck squamous cell carcinoma. Cancer Genet with their profiles of gene expression. Mol Oncol 1: 84–96. Cytogenet 176: 100–106. 28. Coombs LM, Pigott D, Proctor A, Eydmann M, Denner J, et al. (1990) 17. Hinds DA, Kloek AP, Jen M, Chen X, Frazer KA (2006) Common deletions and Simultaneous isolation of DNA, RNA, and antigenic protein exhibiting kinase SNPs are in linkage disequilibrium in the human genome. Nat Genet 38: 82–88. activity from small tumor samples using guanidine isothiocyanate. Anal Biochem 18. Frank B, Bermejo JL, Hemminki K, Sutter C, Wappenschmidt B, et al. (2007) 188: 338–343. Copy number variant in the candidate tumor suppressor gene MTUS1 and 29. Nizak C, Martin-Lluesma S, Moutel S, Roux A, Kreis TE, et al. (2003) familial breast cancer risk. Carcinogenesis 28: 1442–1445. Recombinant antibodies against subcellular fractions used to track endogenous 19. Nouet S, Amzallag N, Li JM, Louis S, Seitz I, et al. (2004) Trans-inactivation of Golgi protein dynamics in vivo. Traffic 4: 739–753. receptor tyrosine kinases by novel AT2 receptor-interacting protein, ATIP. J Biol 30. Dimitrov A, Paupe V, Gueudry C, Sibarita JB, Raposo G, et al. (2009) The gene responsible for Dyggve-Melchior-Clausen syndrome encodes a novel peripheral Chem 279: 28989–28997. membrane protein dynamically associated with the Golgi apparatus. Hum Mol 20. Di Benedetto M, Bie`che I, Deshayes F, Vacher S, Nouet S, et al. (2006) Genet 18: 440–453. Structural organization and expression of human MTUS1, a candidate 8p22 31. Rong R, Jin W, Zhang J, Sheikh MS, Huang Y (2004) Tumor suppressor tumor suppressor gene encoding a family of angiotensin II AT2 receptor- RASSF1A is a microtubule-binding protein that stabilizes microtubules and interacting proteins, ATIP. Gene 380: 127–136. induces G2/M arrest. Oncogene 23: 8216–8230. 21. Wruck CJ, Funke-Kaiser H, Pufe T, Kusserow H, Menk M, et al. (2005) 32. Pirollo KF, Rait A, Zhou Q, Zhang XQ, Zhou J, et al. (2008) Tumor-targeting Regulation of transport of the angiotensin AT2 receptor by a novel membrane- nanocomplex delivery of novel tumor suppressor RB94 chemosensitizes bladder associated Golgi protein. Arterioscler Thromb Vasc Biol 25: 57–64. carcinoma cells in vitro and in vivo. Clin Cancer Res 14: 2190–2198. 22. Reyal F, Stransky N, Bernard-Pierrot I, Vincent-Salomon A, de Rycke Y, et al. 33. Di Benedetto M, Pineau P, Nouet S, Berhouet S, Seitz I, et al. (2006) Mutation (2005) Visualizing chromosomes as transcriptome correlation maps: evidence of analysis of the 8p22 candidate tumor suppressor gene ATIP/MTUS1 in chromosomal domains containing co-expressed genes–a study of 130 invasive hepatocellular carcinoma. Mol Cell Endocrinol 252: 207–215. ductal breast carcinomas. Cancer Res 65: 1376–1383. 34. Aoki K, Taketo MM (2007) Adenomatous polyposis coli (APC): a multi- 23. Rakha EA, El-Sayed ME, Lee AH, Elston CW, Grainge MJ, et al. (2008) functional tumor suppressor gene. J Cell Sci 120: 3327–3335. Prognostic significance of Nottingham histologic grade in invasive breast 35. Ishii H, Vecchione A, Murakumo Y, Baldassarre G, Numata S, et al. (2001) carcinoma. J Clin Oncol 26: 3153–3158. FEZ1/LZTS1 gene at 8p22 suppresses cancer cell growth and regulates mitosis. 24. Vincent-Salomon A, Lucchesi C, Gruel N, Raynal V, Pierron G, et al. (2008) Proc Natl Acad Sci U S A 98: 10374–10379. Integrated genomic and transcriptomic analysis of ductal carcinoma in situ of the 36. Song MS, Song SJ, Ayad NG, Chang JS, Lee JH, et al. (2004) The tumour breast. Clin Cancer Res 14: 1956–1965. suppressor RASSF1A regulates mitosis by inhibiting the APC-Cdc20 complex. 25. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, et al. (2007) Nat Cell Biol 6: 129–137. American Society of Clinical Oncology/College of American Pathologists 37. Warner SL, Gray PJ, Von Hoff DD (2006) Tubulin-associated drug targets: guideline recommendations for human epidermal growth factor receptor 2 Aurora kinases, Polo-like kinases, and others. Semin Oncol 33: 436–448. testing in breast cancer. J Clin Oncol 25: 118–145. 38. Huszar D, Theoclitou ME, Skolnik J, Herbst R (2009) Kinesin motor proteins as targets for cancer therapy. Cancer Metastasis Rev 28: 197–208.

PLoS ONE | www.plosone.org 11 October 2009 | Volume 4 | Issue 10 | e7239